Bill, thanks for the note. And thanks also for
making my point (below).
>>The underlying routing mechanism which
computes the best path through an Ascend
Frame/ATM network is called Virtual Network
Navigator.<<
The mere fact that the IETF protocols ride atop the
ATM fabric means that they are experiencing a
considerable amount of overhead (10% to 20%). That
was one of my main points.
If QoS is indeed crucial, and the budget can stand it,
then depending on other variables, you are probably 100%
correct.
If we're talking about run-of-the-mill IP, or WWW
transport for most applications, then it's a heavier burden to
bear than would exist otherwise, e.g., with generic distributed
routing using Layer 3 routers.
Discussions such as these tend to characterize their
participants [us] as being one way or the other. Just for
the record, I see the benefits of both worlds in their
proper settings. My points were made to question the actual
availability in ASND's gear to support the "native"
IP part, if you read my messages again, and not to
suggest that "pure routed IP" would be the optimum fit
in all cases.
Appreciate the note and the historical facts, tho.
Regards, Frank C.
ps - you say I listened to a CSCO employee? Who would that be?
|
